Conference Report on Stroke Mortality in the Southeastern United States
Abstract—A workshop to describe and then seek possible causes for the increased stroke mortality in the southeastern United States briefly considered 30 suspected correlates and discussed in more detail the 10 thought to be most likely. Recent age-adjusted stroke mortality rates in adults from industrialized countries reveal marked geographic differences. Age-adjusted statewide stroke mortality rates also differ, and they are higher in the Southeast than elsewhere in the United States. For five southeastern coastal states in the heart of the “Stroke Belt,” excess stroke mortality has been present at least since 1930. In a 20-year follow-up of 10 000 veterans, the Stroke Belt had a 25% increase in all-cause mortality and congestive heart failure. A potential cause of increased fatal stroke included hypertension, which was more frequent in the Stroke Belt. No consistent patterns of lifestyle differences or of differences in potassium or calcium intake seemed to explain the higher rates of fatal strokes in the Stroke Belt; however, detailed investigations of smaller populations in localized areas seem warranted. Some data suggest a relationship between socioeconomic status and the Stroke Belt effect. Other differences in the Southeast that could explain, at least partially, the Stroke Belt effect include presence of soft water throughout most of the area, decreased antioxidant intake, and differences in the use of medical care and in the response to antihypertensive drugs. On the basis of available information, the three most likely explanations or partial explanations for the Stroke Belt are increased levels of blood pressure, localized differences in socioeconomic status, and toxic environmental factor(s). Two major recommendations were made: (1) to encourage both patient and caregiver to use all currently available means of decreasing morbidity and mortality by controlling blood pressures at or below normal levels and by reducing other risk factors and (2) to seek precise information about relationships of identified possible causes of increased morbidity and mortality in the Stroke Belt.
Stroke is the third leading cause of death in the United States, and it is significantly more prevalent in some parts of the country than in others.1 Based on 1980 data, 11 states with stroke death rates more than 10% above the average rate for the entire United States form a contiguous cluster in the southeast, with Indiana jutting upward from the cluster.2 This region has come to be called the “Stroke Belt.” Since it was first observed, the excess stroke mortality in the Southeast has aroused the interest of various groups and has been discussed at various meetings; however, the underlying causes for this excess risk remain unexplained. The National Heart, Lung, and Blood Institute and the DVA have had a continuing interest in understanding and then eradicating the excess mortality in the Stroke Belt. To this end a workshop was held in August 1994, with the following objectives: to characterize the Stroke Belt, to examine and discuss its various postulated causes, and to make recommendations about (1) what further information is needed to understand this phenomenon and (2) what steps should be taken now to reduce the associated excess mortality. Special emphasis has been placed on hypertension because of its role as a precursor of stroke.
Stroke Mortality and Suggested Causative Factors
Stroke Mortality Worldwide
Recent age-adjusted stroke mortality rates in middle-aged and elderly adults vary by a factor of 6 to 7 across industrialized countries.3 4 5 6 (Throughout this article, mortality rates are age-adjusted to the 1940 standard and given in terms of deaths per year per 100 000 persons. For national and international mortality, the term “stroke” refers to deaths classified as “cerebrovascular diseases.”) Based on data from 52 countries during the latter part of the 1980s, the age-adjusted annual death rates per 100 000 for persons aged 35 to 74 years were lowest in Switzerland (at 41 in men and 24 in women) and highest in Bulgaria (at 277 in men and 180 in women). Although stroke death rates are lower in women than men, stroke accounts for a larger proportion of deaths in women.
The wide variation in national stroke mortality rates may be partially explained by competing causes of death, differences in lifestyle, local environment, medical care, and inaccuracy of diagnosis or reporting.5 For western Europe the range was approximately twofold in both men and women, varying from the low rates in Switzerland to higher rates in Italy and Spain. Portugal was an outlier with very high rates of 127 in men and 108 in women. The United States and Canada had relatively low rates of 47 and 44, respectively. Except for Poland, which had a rate of 109, the rates were high in Eastern Europe, in the range of 200 to 277. Rates were also high in China and other Asian countries.5
Stroke and All-Causes Mortality in the United States
Not only in 1980 but in 1970 and 1990 as well, the 11 Stroke Belt states in the Southeast had the highest overall stroke mortality rates (Table 1⇓) and separately had the highest rates for men and women and for blacks and whites.7 8 Moreover, except for Arkansas and Louisiana, the other 9 states had the highest stroke mortality rates continuously from 1940 through 1990 (Table 1⇓).
Although all states had markedly higher stroke death rates in 1930 than in 1990,8 age-adjusted stroke mortality rates did not begin to decline markedly in most Stroke Belt states until after 1960, somewhat later than in the rest of the United States. Between 1930 and 1940, only 4 of the 11 Stroke Belt states had steeper declines in stroke mortality rates than the United States as a whole; between 1940 and 1950 only 3 did; and between 1950 and 1960 only 1 did. The gap between the state mortality rates for the entire United States and the generally higher rates in the Stroke Belt appears to have widened modestly over the period from 1930 to 1990. For the total population, stroke mortality in 1990 averaged 33% higher in the Stroke Belt states than in the other states. The corresponding percentages were higher for black men (40%) than for white men (23%) and higher for black women (38%) than for white women (16%).8
In recent but not in earlier years, the Stroke Belt states ranked high in all-causes mortality as well as in stroke mortality. The correlation coefficient of the ranking for all-causes and for stroke mortality rates for the 50 states was .65 in 1980 but only .14 in 1950.7 Thus, all-causes mortality has come to be highly related to stroke mortality.
15-Year All-Causes Mortality and Cardiovascular Morbidity Among Veterans
In the mid-1970s the DVA HSTP identified 11 912 previously untreated hypertensive patients whose baseline and early treatment data were collected and computerized.9 During the subsequent 15 years, 45% (5337) of these veterans died. Their dates of death and of major cardiovascular morbidity (stroke, myocardial infarction, congestive heart failure, and end-stage renal disease) were obtained from DVA records and the National Death Index.10
The mortality of patients treated at HSTP clinics in southeastern states was compared with the mortality of their counterpart patients in nine western states.11 Six HSTP clinics (Birmingham, Indianapolis, Jackson, Memphis, New Orleans, and Richmond) were located in the Stroke Belt. Four additional HSTP clinics were located in Washington, DC, Texas, and Oklahoma, areas that also report higher age-adjusted stroke mortality than the rest of the nation. The data from these four clinics have been added to the data from the six clinics in the Stroke Belt proper, thus providing greater statistical power. These 10 clinics have been included in an “Expanded Stroke Belt.” All-causes mortality for patients in the Expanded Stroke Belt was then compared with the mortality for the 10 HSTP clinics in nine non–Stroke Belt states west of the Mississippi River (Figure 1⇓).
Patients treated in HSTP clinics in the Expanded Stroke Belt were approximately the same age (52.6±10.3 years) as their counterparts treated in the nine western non–Stroke Belt states (52.8±10.5 years), but the former were more likely to be black (57.1% versus 24.7%) and had higher average pretreatment systolic (154.5±19.9 versus 151.8±15.3 mm Hg) and diastolic (100.7±10.1 versus 98.9±8.8 mm Hg) blood pressures. Multivariate proportional hazards models that controlled for age, blood pressure, and race revealed a significant elevation in risk ratio for all-causes mortality in the Stroke Belt proper of 1.217 and in the Expanded Stroke Belt of 1.259 versus the non–Stroke Belt (Table 2⇓).
Thus, both the Stroke Belt proper and the Expanded Stroke Belt have increased rates of all-causes mortality compared with a non–Stroke Belt region of the United States. In addition, the Expanded Stroke Belt has increased rates of DVA hospitalizations for end-stage renal disease, congestive heart failure, and stroke (Table 2⇑).
US National Data on High Blood Pressure
NHANES III (1988 to 1991) found the expected progressive increase in average blood pressure with increasing age for both men and women.12 The age-adjusted average blood pressures for non-Hispanic blacks were generally higher than those for non-Hispanic whites or for Mexican Americans. The age-adjusted values for prevalence of hypertension in these three groups were 32.4%, 23.3%, and 22.6%, respectively.12 There are no satisfactory data that compare the prevalence of hypertension in the Stroke Belt states with the prevalence in other states.
Awareness, treatment, and control of hypertension also varied by age, race, and gender.12 The pattern for these parameters was similar to the “prevalence pattern” described above. Not surprisingly, for the race-gender groups listed above, treated hypertensives had lower average systolic and diastolic blood pressures than untreated hypertensives; however, the average blood pressures of treated hypertensives was still 10 to 20 mm Hg higher than for their normotensive counterparts.
Finally, the four successive national health surveys conducted from 1960 to 1964 and from 1988 to 1991 suggest a progressive decline in average blood pressure and in the prevalence of hypertension.13
Other Conditions That Might Increase Stroke Rates
Diabetes mellitus is a potent accelerator of atherosclerosis and therefore of myocardial infarction, stroke, and peripheral vascular disease.14 Diabetes is more frequent in the presence of obesity, a particular problem for black women.
Hypertension, itself the most important risk factor for stroke, and diabetes mellitus are each responsible for approximately one third of the cases of chronic renal failure. The added presence of the characteristic lipid abnormalities observed with some chronic renal disease would seem likely to increase the number of fatal strokes.
Although atrial fibrillation with associated atrial thrombus formation can lead to brain emboli, data on geographic differences in its frequency are limited. Rheumatic valvular disease is now infrequent, but nonvalvular atrial fibrillation usually occurs after age 65 and should be examined as possibly being important in excess stroke mortality of the elderly in the Southeast.
Lifestyle Characteristics Leading to Hypertension and Stroke
National Health Interview Survey data from 1991 suggest that the fraction of both men and women who use salt, are overweight, report no leisure time physical activity, and smoke cigarettes was generally highest for the Southeast compared with the other regions of the United States.15 In contrast, alcohol intake was lowest for the Southeast compared with the other regions.16
When risk factor data from the National Health Interview Survey for 1990 and the NHANES for 1988 to 1991 were analyzed by four age groups (18–29, 30–44, 45–64, ≥65 years), the pattern was less clear.17 The fraction of persons 20% or more above desirable body weight was highest in the Southeast among the two younger age groups but lowest among the two older age groups for both men and women. Southern black women were more likely to be overweight than the other race-gender groups. Except for the youngest age group, the percentage of men currently smoking cigarettes was highest for the Southeast; there was no clear pattern for women. Daily consumption of salt intake by 24-hour recall was highest in the Southeast, except for black men, for whom the converse was true; in general, men consumed approximately 1000 mg/d more than women. The percentage of both men and women who had at least one drink of alcohol in the past year was lowest for the Southeast for all four age groups, and the difference was most marked for the two older groups. There were no clear regional differences in the number of persons getting regular exercise. The fraction of persons with self-reported adverse health effects due to stress was lowest in the Southeast for both men and women, especially among the three younger age groups.
Thus, there are regional differences in many of the usual risk factors for hypertension and its complications16,17; however, with the available data, it is difficult to find a consistent pattern that might explain the persistent Stroke Belt, although a detailed look at smaller population groups and more localized geographic areas might be helpful.
Dietary Potassium and Calcium in the Stroke Belt
The observation of high stroke rates persisting for 50 years in a large geographic area is consistent with an environmental common-source epidemic. Two possible sources of exposure are (1) a specific dietary nutrient and (2) a potential toxic chemical in water or soil. The increasingly general distribution of food throughout the country lessens the likelihood that a specific food contaminant is the culprit.
The “Three Area Study” in the 1970s did not demonstrate that whites living in the South had higher blood pressures than whites in other communities. The only risk factor for stroke that appeared to consistently differ by area and race was the prevalence of elevated blood glucose.18
Potassium intake is lower in blacks with low SES.19 It is also lower in the Southeast than elsewhere. Although the suggestion of an inverse relationship between dietary potassium and low blood pressure is gaining increased acceptance, the relationship is far from proven. Potassium may, however, have a direct effect on stroke independent of blood pressure level. A testable hypothesis is that diet (ie, high salt intake, excess caloric intake leading to obesity, and low potassium intake) accounts for the higher rate of stroke among blacks and generally among the population residing in the Stroke Belt.
Clinical trials of dietary change among groups with low SES and among blacks need to be done. In the interim, it may be helpful to increase potassium intake; traditional risk factors, ie, smoking, obesity, blood lipids, and high blood pressure, should certainly be decreased.
Socioeconomic Status and Stroke Mortality in the Southeast
Stroke mortality is hypothesized to be associated with low SES of individuals and of the social environment of their place of residence. High blood pressure, smoking, and diabetes have been demonstrated to be stroke risk factors in cohort studies in the South in both whites and blacks and in both men and women.21 Among blacks in Evans County, Georgia, age-adjusted stroke and heart disease mortality risks were associated with education and SES, with higher disease rates among those with lower SES and those who are less educated. In addition, the extent of preclinical carotid atherosclerosis varies inversely with SES. Aggregate levels of SES have been lower in the Southeast than the rest of the nation for the past 30 years.
Age-adjusted risk ratios for stroke mortality in the Evans County 30-year follow-up were 1.4 and 2.6 for white and black women, respectively, with less than a high school education compared with white women who had more than a high school education.22
In the Hypertension Detection and Follow-up Program, 5-year age-race-gender–adjusted mortality rates increased stepwise with decreasing education in hypertensives referred to usual care. Hypertensives referred to stepped care had 19% lower overall mortality than usual-care patients, and the differences in mortality associated with education disappeared.23
The prevalence of hypertension, its severity, and its comorbidity varied inversely with SES in the Atherosclerosis Risk in Communities Study.24 Preclinical carotid atherosclerosis increased with increasingly severe hypertension. However, atherosclerosis varied inversely with SES at each stage of hypertension.
In 1960, overall stroke mortality rates were higher in the Southeast than in the rest of the nation, and they were highest among residents of regions characterized by low educational achievement and low average income and occupational levels. There was no excess of stroke mortality in the Southeast among residents of areas with the highest educational levels.
Some vital statistics data suggest that the mortality differences between the Southeast and the remainder of the country have decreased since 1960, an interval during which there have been marked declines in mortality from stroke, coronary heart disease, and all cardiovascular disease, as well as total mortality. Although the validity of the data is uncertain, higher mortality rates for deaths attributed to stroke in the Southeast are clustered in places of residence characterized by low SES. It is possible that there are areas where medical treatment is inadequate. Thus, observational studies and clinical trials found inverse associations of stroke with SES, and currently a residual excess stroke mortality in the Stroke Belt persists in areas characterized by a high proportion of residents with low educational achievement.
Genetic Approach to Hypertension
Genetic approaches to understanding the pathophysiology of complex human traits, such as hypertension, can complement physiological analyses and should improve our ability to treat or prevent the disease; however, the currently available information is too limited to throw any light on the mechanism of the Stroke Belt. Linkage analysis with candidate genes with the use of intermediate phenotypes has identified two genes involved in hypertension.25 26
The first is a fusion gene mutation linking the regulatory region of the 11-hydroxylase gene to the coding sequence for aldosterone synthetase.25 This mutant gene is responsible for glucocorticoid-remediable aldosteronism. The intermediate phenotype used in studies of the gene was an increase in levels of the adrenal steroids 18-oxycortisol and hydroxycortisol. The gene for glucocorticoid-remediable aldosteronism was identified by means of a pedigree approach, a method that is likely to identify other genes involved in hypertension. The most appropriate population for studies of genetic predictors of hypertension would be affected sibling pairs who both have hypertension.27
In a recent study the angiotensinogen gene was also linked to hypertension in individuals who had severe or early-onset hypertension. A variant of the angiotensinogen gene with threonine rather than methionine at codon 235 was specifically associated with hypertension. The T235 homozygote of the angiotensinogen gene was associated with the nonmodulating intermediate phenotype of essential hypertension. Since angiotensin-converting enzyme inhibitors appear to correct the specific defect underlying the elevated blood pressure in nonmodulators, identification of the gene potentially associated with nonmodulation raises the strong possibility that genetic screening will permit more specific therapy. This promising field has the potential to provide a clearer understanding of hypertension; however, further information is needed to reach this potential.
Differing Responses to Antihypertensive Drugs in the Stroke Belt
The antihypertensive efficacy of six drugs and a placebo has recently been compared in a randomized double-blind study involving 1292 male veterans with previously untreated hypertension.28 Fifteen DVA clinics were included in the study, with six of them being from the southeastern United States, ie, the Stroke Belt region.
In a nonrandomized comparison, antihypertensive drug efficacy in Stroke Belt clinics was compared with efficacy in non–Stroke Belt clinics.29 For hydrochlorothiazide, atenolol, clonidine, and captopril, patients in the Stroke Belt achieved successful blood pressure control less frequently than non–Stroke Belt patients; for diltiazem and prazosin, there were no obvious differences.
Examining the data for the two races separately suggested that blacks were better able to control their blood pressure with diltiazem than whites but less able to do so with prazosin. Blacks were also less likely than whites to achieve satisfactory blood pressure control with atenolol. Finally, in the Stroke Belt, captopril appeared less effective in controlling blood pressure in blacks than in whites. The reasons for these differences in efficacy are unclear. Regression analysis suggests that they are not explained by income level, medication compliance, sodium or potassium intake, race, or blood renin levels.
Possible Environmental Toxicity in the Stroke Belt
The increased stroke mortality rates in a large geographic region for a long period of time raise the possibility of environmental toxins. Two likely sources of such toxins are (1) water and (2) soil.
Soft water has frequently been associated with increased cardiovascular mortality, particularly stroke and myocardial infarction, since 1960 when Schroeder30 first reported the correlation. His initial report was based on 1950 census data in the United States; the 1960 census data confirmed this observation. Moreover, Morris31 found the same relationship in Great Britain, using both 1950 and 1960 data. Since then there have been many reports relating soft water to increased cardiovascular mortality, although extensive efforts have failed to pinpoint a specific mechanism for the effect.32
The coastal states of the Stroke Belt constitute the largest soft water area in the United States. Almost the entire states of North Carolina, South Carolina, Georgia, Alabama, Mississippi, Arkansas, and Louisiana (except for a 100-mile strip along the Mississippi River) have very soft surface water with less than 60 ppm of calcium carbonate. Northern Florida, southern Tennessee, eastern Texas, and eastern Virginia have equally soft water. Kentucky and the rest of Virginia have somewhat harder but still soft water (Figure 2⇓). The fact that the coastal Southeast has very soft water and has had very high stroke mortality rates for more than 50 years suggests the possibility that water hardness—or some associated characteristic—may help to explain the high stroke rate in the Stroke Belt. However, other places with lower stroke mortality rates also have soft water, in particular New England and the Pacific Northwest. On the other hand, Indiana, which has a high stroke mortality rate, has harder water. Thus, soft water could play a part in elevated stroke mortality rates but is not the entire explanation.
Among other possible environmental toxicities, the most likely would seem to be exposure to lead or perhaps cadmium. The presence of lead, even at low blood levels, has long been associated with hypertension and also with stroke.33 The limited available data, however, can neither include nor exclude low environmental (as opposed to higher occupational) levels of lead as a significant contributor to human hypertension.34 Moreover, unlike the soft water situation, there are no reliable data that indicate exposure of the entire population to lead in various regions of the United States; thus, there is no convincing evidence of excess lead exposure in the Stroke Belt. In support of a possible lead effect, long-term exposure of rats to as little as 1 ppm of lead in drinking water regularly induces a persistent 10 to 15 mm Hg increase in systolic blood pressure,35 mimicking both human essential hypertension and common tissue lead levels. Thus, a relatively small blood pressure effect could be present in a significant part of a regional population and could explain an increased stroke rate.
Like lead, very-low-level cadmium exposure can induce hypertension in animals36 and has been suspected of inducing it in humans. Ingested cadmium is permanently sequestered in the kidney, but it has proved technically difficult to demonstrate increased renal cadmium levels as a possible cause of early, uncomplicated hypertension. Cadmium is particularly available as a frequent contaminant of sludge that is used as fertilizer, but the extent to which this introduces cadmium into the food chain is uncertain, and again there is no evidence of increased exposure in the Southeast.
Another possibility is lack of a beneficial substance rather than an excess of a toxic substance. The absence of antioxidants has been suggested as a possible contributor to strokes; a dearth of selenium has received considerable attention,37 but there are even fewer data associating such substances with protection against cardiovascular disease than there are for toxicity as a contributor to cardiovascular disease.
Thus, the very soft water of the southeastern coastal states could contribute to the persistent high stroke mortality rate in the Stroke Belt. Of the metals that have been considered, low-level lead seems the most likely culprit, although increased exposure in the Stroke Belt remains to be shown.
The CARDIA Study and the Stroke Belt
The CARDIA Study was initiated in 1985 to examine the distribution and evolution of risk factors for coronary artery disease in young adult black and white men and women.38 Over 5000 participants aged 18 to 30 years have been followed for 7 years in four American cities: Birmingham, Ala; Chicago, Ill; Minneapolis, Minn; and Oakland, Calif. These participants have had four study examinations that assessed traditional risk factors (eg, blood pressure, lipids, smoking), family history of cardiovascular disease, educational level, dietary factors, and left ventricular mass by echocardiogram.38
Levels of selected factors and, when appropriate, changes in these factors over 7 years were compared between Birmingham, a Stroke Belt city, and the other three centers. In Birmingham, the change in diastolic blood pressure during follow-up was greater in all four race-gender groups than it was in the three non–Stroke Belt cities; the change in systolic blood pressure was greater only in blacks.39 In Birmingham, the prevalence of hypertension (>140/90 mm Hg or on medication) was higher in all but white women, potassium excretion was lower in all groups (Table 3⇓), and consumption of beta carotene and nonstarchy vegetables was lower in all groups. Consumption of fruits was lower in whites, and fibrinogen levels were slightly higher in whites. Birmingham did not have more adverse levels of risk factors for the following: baseline body mass index, weight change, smoking, alcohol consumption, LDL cholesterol, insulin level, prevalence of migraine headache, baseline systolic or diastolic blood pressure, urinary excretion of sodium or magnesium, intake of calorie-adjusted total or polyunsaturated fat, or left ventricular mass. In several instances, levels of factors in Birmingham were more favorable than in other centers: lower baseline body mass index and systolic and diastolic blood pressure among white women; lower rates of smoking, lower sodium excretion, and lower left ventricular mass among black women; and lower intake of saturated fat in all groups.39
It is known that the higher rates of hypertension contribute to the higher risk of stroke in the Southeast, and it is possible that lower intake of fruits, vegetables, and antioxidants in the Southeast contributes to the greater rise in blood pressure with age.40
Use of Office-Based Medical Care: South Carolina Data
The age-adjusted prevalence of hypertension in South Carolina (140/90 mm Hg and/or pharmacological antihypertensive treatment) is higher among blacks than among whites (36.9% in black men versus 29.4% in white men; 39.6% in black women versus 23.0% in white women). There are similar variations in the percentages treated and controlled.
The South Carolina Ambulatory Medical Care Survey,41 which assessed patient visits for office-based physicians in a manner similar to that used in the National Ambulatory Medical Care Survey,42 noted that whites utilized office-based medical care services at greater rates than blacks in both the South Carolina Survey and the National Survey. Thus, in South Carolina, white men averaged 2.45 and black men averaged 1.50 visits per year. Similarly, white and black women averaged 3.73 and 2.30 visits per year, respectively. Frequency of patient visits also varied by mode of payment and by rural versus urban residence. Racial discrepancies were greater for visits to specialists than for visits to generalists. Blood pressure assessments were incorporated into the patient visits at similar rates for both race groups.
These presentations and the subsequent discussions engendered an extended list of factors that were considered possible contributors to the Stroke Belt effect (Table 4⇓), and recommendations for future action were proposed.
Search for Causes of Excess Mortality
A major recommendation was to seek the information needed to better understand the Stroke Belt phenomenon. Understanding will require improved and reliable prevalence information on parameters potentially related to stroke mortality. A serious effort to understand may involve studies specifically designed to compare such parameters in Stroke Belt and non–Stroke Belt states. Of the various possible explanations for the Stroke Belt, three seemed most likely on the basis of presently available information and hence were felt to deserve extensive exploration: (1) increased mortality among persons with low SES, (2) environmental toxicity, and (3) the effects of hypertension alone or in conjunction with some combination of the usual recognized risk factors for stroke.
To explore the first requires careful study of specific high-risk subpopulations in small geographic areas within the individual states to determine the extent to which excess stroke mortality is concentrated in particular groups of individuals from the lowest SES category. Such a study should attempt to relate SES of the target populations to blood pressure as well as to other risk factors.
To explore environmental toxicity involves examining past and present patterns of possible toxic influences within the Stroke Belt area, again emphasizing both geographic and population details. Potentially toxic environmental substances that merit consideration include but are not limited to hardness/softness of water (ground, surface, and finished water) and the trace elements (lead and cadmium exposure and lack of dietary selenium). Consideration should also be given to nonenvironmental toxicities that might be different in the Stroke Belt, eg, analgesic use or even hantavirus infection.
To explore the effect of recognized risk factors requires a search for excesses of these risk factors, most likely in small defined populations in small defined geographic areas. This should include obtaining representative blood pressure data by age, race, and gender for the Stroke Belt compared with other regions in the United States.
Other topics that deserve investigation include the following: (1) the westward extension of the Stroke Belt during the last 50 years, first to include Arkansas and Louisiana and later to include Texas and Oklahoma; (2) the relationship of stroke mortality to total mortality and its change with time; (3) the relationship between stroke mortality and mortality due to myocardial infarction, congestive heart failure, and end-stage renal disease. The incidence of nonfatal cases of these cardiovascular conditions should also be explored since the ratio of fatal to nonfatal cases is decreasing. Noncardiovascular causes of death should also be investigated since some seem to be increased in the Stroke Belt, eg, carcinoma of the lung; and (4) the question of whether the Stroke Belt phenomenon is increasing or decreasing. It was recommended that these research topics be intensively studied and that the new information obtained be considered in depth in some appropriate forum.
An additional general recommendation, which received strong consensus support, stated that since the Southeast is a unique high-mortality area, the causes and relationships of the Stroke Belt should be studied as long as the phenomenon persists, even while strenuous efforts are being made to eliminate excess mortality.
Education to Improve Patient Care
A major recommendation was to educate individual primary healthcare providers, in particular providers for patients with low SES and patients from rural areas. These providers are usually family physicians, physician assistants, or registered nurses. The useful treatment information that should be emphasized includes (1) an understanding of the excess risks associated with the Stroke Belt; (2) the importance of treating and of controlling hypertension, particularly in high-risk subpopulations, with cost-effective regimens; and (3) the value of an initial trial of low-dose diuretic or diuretic plus β-blocker with the use of more expensive angiotensin-converting enzyme inhibitors and calcium blockers being reserved for special circumstances or when necessary.43 This may require identifying high-risk individuals by targeted screening. The high-risk group certainly includes patients with elevated systolic as well as diastolic pressures, ie, with isolated or primarily systolic hypertension. The educational program will have to be developed at professional education meetings, in special societies being formed in the Stroke Belt area (eg, Consortium of Southeastern Hypertension Centers), at DVA clinics that are offering to provide preceptorships, and by journal articles and editorials.
The other aspect of the recommended educational effort involves community workers who should be instructed in optimum methods to utilize their limited resources and at the same time be provided information on the expected results of various approaches. In this context, the epidemiology of cardiovascular morbidity and mortality should be presented and their prevention should be emphasized.
There is a Stroke Belt; however, its magnitude is somewhat uncertain since it depends on stroke mortality obtained from death certificate data. There is also an increased total mortality in the Southeast that is not subject to the uncertainty of cause-specific mortality. In addition, there is an increase in end-stage renal disease, which may well be related to the same or similar underlying factor(s).
The exact cause of the Stroke Belt is unknown, but possible explanations have been identified. The level of blood pressure and the prevalence and severity of hypertension are higher in the Southeast than elsewhere, although the reasons for this are not clear. How much this increase in blood pressure explains the observed increase in stroke rate is unclear; however, small changes in blood pressure do have large effects on a population’s cardiovascular disease. The data on other risk factors do not suggest that any of them make a significant contribution to the explanation of increased stroke mortality, although adequately detailed information is not available.
Low SES of the stroke victims may be responsible for a portion of the increased Stroke Belt mortality. However, much more detailed data on local populations in local areas are needed to explain the observed increase in mortality. It is known that low SES is associated with increased risk for many diseases, including cardiovascular diseases.
The possibility that environmental toxicity of some unspecified sort is responsible for the Stroke Belt is consistent with the long and persistent occurrence of the Stroke Belt phenomenon. It is possible that soft water, which has been found to be associated with increased cardiovascular mortality and morbidity, is involved; however, whether and how much soft water contributes to excessive stroke mortality are unknown. The extent to which excesses of toxic substances like lead or cadmium are involved as precursors of hypertension cannot be adequately evaluated with the currently available data.
Obviously, more precise information is needed to explain the riddle of the Stroke Belt. Whatever the cause of the Stroke Belt, prompt and effective efforts to control hypertension, which is a major risk factor for stroke, are urgently needed.
Selected Abbreviations and Acronyms
|DVA||=||Department of Veterans Affairs|
|CARDIA||=||Coronary Artery Risk Development in Young Adults|
|HSTP||=||Hypertension Screening and Treatment Program|
|NHANES||=||National Health and Nutrition Examination Survey|
Participants in the Workshop
National Heart Lung and Blood Institute, Bethesda, Md: Diane E. Bild, Jeffrey Cutler, Millicent Higgins, Michael Horan, Phyliss Sholinsky, Thomas Thom; DVA and Washington University School of Medicine, St Louis, Mo: H. Mitchell Perry, Jr; DVA Medical Center, Memphis, Tenn: William Cushman; University of Vermont, Burlington: Harriet Dustan; National Center for Health Statistics, Hyattsville, Md: Manning Feinleib, Lillian Ingster; Bowman Grey School of Medicine, Winston-Salem, NC: Carlos Ferrario; Hames Clinic, Claxton, Ga: Curtis Hames; University of Pittsburgh (Pa): Lewis Kuller; Medical University of South Carolina, Charleston: Daniel Lackland; Washington University School of Medicine, St Louis, Mo: Philip Miller; University of Maryland Hospital, Baltimore: Thomas Price; University of North Carolina, Chapel Hill: Herman Tyroler; Johns Hopkins University School of Hygiene and Public Health and School of Medicine, Baltimore, Md: Paul Whelton; Brigham and Women’s Hospital, Boston, Mass: Gordon Williams.
The Workshop on Stroke Mortality in the Southeastern United States was held by the National Heart, Lung, and Blood Institute and DVA in Bethesda, Md, August 29–30, 1994, and was supported by the National Heart, Lung, and Blood Institute and the DVA.
Reprint requests to Dr H. Mitchell Perry, Veterans Affairs Medical Center (111DJC), 915 N Grand Blvd, St Louis, MO 63106.
- Received September 12, 1997.
- Revision received October 17, 1997.
- Accepted January 8, 1998.
Roccella EJ, Lenfant C. Regional and racial differences among stroke victims in the United States. Clin Cardiol. 1989;12:18–22.
National Heart, Lung, and Blood Institute. Data Fact Sheet: The Stroke Belt: Stroke Mortality by Race and Sex. Hyattsville, Md: National Institutes of Health, US Dept of Health and Human Services; 1989.
World Health Statistics Annual. Geneva, Switzerland: World Health Organization; 1988.
World Health Statistics Annual. Geneva, Switzerland: World Health Organization; 1991.
World Health Statistics Annual. Geneva, Switzerland: World Health Organization; 1992.
World Health Statistics Annual. Geneva, Switzerland: World Health Organization; 1993.
Lanska DJ. Geographical distribution of stroke mortality in the United States: 1939–41 to 1979–81. Neurology. 1993;43:1839–1851.
Lanska DJ, Peterson PM. Geographic variation in the decline of stroke mortality in the United States. Stroke. 1995;26:1159–1165.
Perry HM, Miller JP, Fornoff JR, Baty JD, Sambhi MP, Rutan G, Moskowitz DW, Carmody SE. Early predictors of 15-year end-stage renal disease in hypertensive patients. Hypertension. 1995;25:305–313.
Miller JP, Perry HM Jr, Rossiter JE, Baty JD, Carmody SE, Sambhi MP. Regional differences in mortality during 15-year follow-up. Hypertension. 1994;23:431–438.
Burt VL, Whelton P, Roccella EJ, Brown C, Cutler JA, Higgins M, Horan MJ, Labarthe D. Prevalence of hypertension in the US adult population: results from the Third National Health and Nutrition Examination Survey, 1988–1991. Hypertension. 1995;25:305–313.
Burt VL, Cutler JA, Higgins M, Horan MJ, Labarthe D, Whelton P, Brown C, Roccella EJ. Trends in the prevalence, awareness, treatment, and control of hypertension in the adult US population: data from the Health Examination Surveys, 1960 to 1991. Hypertension. 1995;26:60–69.
National High Blood Pressure Education Program Working Group. National High Blood Pressure Education Program Working Group on hypertension and diabetes. Hypertension. 1994;23:145–158.
Adams PF, Benson V. Current estimates from the National Health Interview Survey, 1991. National Center for Health Statistics. Vital Health Stat 10 (184); 1991.
Palmer AJ, Fletcher AE, Bulpitt CJ, Beevers DG, Coles EC, Ledingham JG, Petrie JC, Webster J, Dollery CT. Alcohol intake and cardiovascular mortality in hypertensive patients: report from the Department of Health Hypertension Care Community Project. J Hypertens. 1995;12:957–964.
Adams PF, Benson V. Current estimates from the National Health Interview Survey, 1990. National Center for Health Statistics. Vital Health Stat 10 (181); 1991.
Stolley PD, Kuller LH, Nefzger MD, Tonascia S, Lilienfeld AM, Miller GD, Diamond EL. Three-area epidemiological study of geographic differences in stroke mortality, II: results. Stroke. 1977;8:551–557.
Langford HG, Stamler J, Wassertheil-Smoller S, Prineas RJ. All-cause mortality in the Hypertension Detection and Follow-up Program: findings for the whole cohort and for persons with less severe hypertension, with and without other traits related to risk of mortality. Prog Cardiovasc Dis. 1986;24:29–54.
Jackson R, Chambless LE, Yang K, Byrne T, Watson R, Folsom A, Shahar E, Kalsbeek W, for the Atherosclerosis Risk in Communities (ARIC) Study Investigators. Differences between respondents and nonrespondents in a multicenter community-based study vary by gender ethnicity. J Clin Epidemiol. 1996;12:1441–1146.
Hopkins PN, Lifton RP, Hollenberg NK, Jeunemaitre X, Hallouin MC, Skuppin J, Williams CS, Dluhy RG, Lalouel JM, Williams RR, Williams GH. Blunted renal vascular response to angiotensin II is associated with a common variant of the angiotensinogen gene and obesity. J Hypertens. 1996;14:199–207.
Materson BJ, Reda DJ, Cushman WC, Massie BM, Freis ED, Kichar MS, Hamburger RJ, Fye C, Lakshman R, Gottdiener J, Ramirez EA, Henderson WG, for the Department of Veterans Affairs Cooperative Study Group on Antihypertensive Agents. Single-drug therapy for hypertension in men. N Engl J Med. 1993;328:914–921.
Cushman WC, Reda DJ, Materson BJ, Perry HM, for the Department of Veterans Affairs Cooperative Study Group on Antihypertensive Agents. Stroke Belt as a predictor of antihypertensive efficacy with single-drug therapy. Am J Hypertens. 1995;8:34A. Abstract.
Perry HM, Perry EF. Cardiovascular disease: the effect of hard and soft water. In: Skandia International Symposia on Trace Elements in Health and Disease. Stockholm, Sweden: Nordiska Bokhandelns Förlag; 1985:193–216.
Pirkle JL, Schwartz J, Landis JR, Harlan WR. The relationship between blood lead levels and blood pressure and its cardiovascular risk implications. Am J Epidemiol. 1985;121:246–258.
Perry HM, Erlanger MW, Perry EF. Elevated systolic pressure following chronic low-level cadmium feeding. Am J Physiol. 1977;232:H114–H121.
Kiefe CI, Williams OD, Bild DE, Lewis CE, Hilner JE, Oberman A. Regional disparities in the incidence of elevated blood pressure among young adults: the CARDIA Study. Circulation.. 1997;96:1082–1088.
Lackland D, Egan B, Dunbar J, Cheek D. Racial differences in the use of office-based medical services for blood pressure assessment. Am J Hypertens. 1993;6:88A. Abstract.
National Ambulatory Medical Care Survey—1991 summary. National Center for Health Statistics. Vital Health Stat 13 (116); 1994.